A218 L13 by CHUA QIU LING

Fundamental

Pengiunieee — 2023-02-09 04:47:52 UTC

MSA
- Experimental
- Mathematical

Use to determine: variation and measurement contributing to overall process

Variation source

Pengiunieee — 2023-02-09 04:53:29 UTC

Sample usage
Equipment used
Operator
Protocol
Environment

Types of Gage Repeatability & Reproducibility (GR & R) Study

20029158 — 2023-02-09 05:04:52 UTC

Type of GR&R Study: Crossed Design

20029158 — 2023-02-09 05:11:52 UTC

Crossed Design is used since each of the 9 nozzles were measured twice by 2 operators in random orders to generate 36 measurements

Two-way ANOVA

20029158 — 2023-02-09 05:13:13 UTC

Based on the p-value, since p-value for nozzle (0.00) < alpha value (0.05) is shows that both contribute to a statistically significant amount of variation to the measured nozzle.
Even though p-value of operator (0.05) = alpha value (0.05), p-value should be higher than 0.05 to be justified as not significant in amount of variation observed.
However, since p-value for nozzle is smaller than operator, it suggests that the characteristics of nozzle may be causing operator to measure particular nozzles differently.

Based on the p-value from interaction between nozzle and operator (0.707) > alpha value (0.05), it indicates that the interaction between nozzle and operator is not statistically significant

Gage R & R (Variance Components)

20029158 — 2023-02-09 05:13:46 UTC

Based on the Total Gage R&R % contribution that should be as small as possible, since it is at 0.80% < 1%, measurement system is acceptable based on guidelines from AIAG whereby most variation observed for the measurement of nozzle is caused by the nozzles and not the equipment of QC engineers

Based on the subcategories of Total Gage R&R, repeatability contributes towards 0.69 % and reproducibility (operator) contributes towards 0.11% of total % contribution, it indicates that the variation from repeated nozzle measurements by same QC engineer is higher than the variation when nozzle is measured by different QC engineers

Based on the % contribution for Part-To-Part, it indicates that the nozzle measurement captured close to 100% of the part variation, and since % is high, it also indicates that system can distinguish between nozzles

Gage Evaluation

20029158 — 2023-02-09 05:14:21 UTC

Based on the % Study Var that is 8.97%, < 10%, the measurement is acceptable based on the guidelines from AIAG.

Based on % Tolerance & % Process that is 4.63% and 9.38% respectively, < 10% the measurement is acceptable based on the guidelines from AIAG

Based on the Number of Distinct Catergories(NDC) = 15, more than the AIAG recommended value of 5 or more, it denotes an acceptable measurement system

Components of Variation

20029158 — 2023-02-09 05:16:19 UTC

Based on the graphical representation that allow quick comparison between contributions of part - to - part and measurement system variation to total variation.
This graph shows that most of the variations is caused by the part-to-part variation since it has the highest percentage out of all the different components

Gage R & R (ANOVA) Report

20029158 — 2023-02-09 05:16:32 UTC

R Chart (Operator Consistency)

20029158 — 2023-02-09 05:17:07 UTC

Based on the comparison between the nozzle measurement done by QC engineer 1 and 2, QC engineer 2's measurement showed more variability in comparison as it as more points that touches the LCL, set control limits

Xbar Chart (Compare Part-To-Part variation to repeatability component)

20029158 — 2023-02-09 05:17:58 UTC

Most of the points that present the mean diameter of nozzle measured in X bar chart are statistically out of control, with only Nozzle 4 being within the control limits set.
This indicates the variation between the measured nozzles (part-to-part variation) across the QC engineers is greater than expected

Measurements by Parts

20029158 — 2023-02-09 05:19:51 UTC

Based on graph, the there is no significant variability in measurements for nozzle, as all the points are close and overlapping the average measurement.
This means that the QC engineers had no difficulty in measuring the nozzle precisely

Measurements by Operators

20029158 — 2023-02-09 05:20:15 UTC

Comparing nozzle measurements done by QC engineer 1 and 2, measurements done by QC engineer 2 appear slightly lower than QC engineer 1

Measurements by Operator and Part

20029158 — 2023-02-09 05:21:17 UTC

Since there is no significant difference nozzle measurement done as most points are exactly overlapping each other, it indicates that there is no disagreement between the nozzle measurements done by the QC engineers.

Pengiunieee — 2023-02-09 05:24:33 UTC

Precision
- repeatability
- reproducibility
Accuracy
- Stability
- Bias - deviation of average and reference
- Linear

Conclusion

20029158 — 2023-02-09 05:29:06 UTC

Since all the values generated from GR&R Study (Crossed) suggests that the measurement system is acceptable based on the guidelines from AIAG, the MSA is acceptable

Crossed Design

20029158 — 2023-02-09 05:31:46 UTC

- Used when each part is measured multiple times by each operator
- Used for non-destructive scenarios

Nested Design

20029158 — 2023-02-09 05:33:43 UTC

- Used when each part is measured by 1 operator
- Used in destructive testing scenarios (batch is assumed to be homogenous)

Root Causes for Poor GR & R Studies

20029158 — 2023-02-09 05:35:23 UTC

Larger Reproducibility

20029158 — 2023-02-09 05:36:38 UTC

- measurement procedures NOT clear
- operator NOT properly trained in using and reading gauge
- operational definition unclear -> NO specific measure as such everyone collects data the same way

Larger Repeatability

20029158 — 2023-02-09 05:36:55 UTC

- lack of maintenance for equipment
- human error
: environment factors
: physical conditions

Biasness

Pengiunieee — 2023-02-09 06:23:11 UTC

Bias = "Bias" / "Tolerance or Process variation" x 100%

If bias is high: possible root causes
- not following measurement protocol
- measurement error in the reference value
- instability

Precision

Pengiunieee — 2023-02-09 06:51:20 UTC

Repeatability
- one measurement variation when done by 1 operator

Reproducibility
- conducted by different operators

Measurement System Analysis (MSA)

20029158 — 2023-02-09 07:24:30 UTC

Stability

Pengiunieee — 2023-02-09 07:32:12 UTC

Linearity

Pengiunieee — 2023-02-09 07:34:02 UTC

Formula

Pengiunieee — 2023-02-09 07:36:22 UTC

% Linearity = "Linearity " /"Process Variation " x 100%

Bias= [slope *Reference Value] + constant

(Y= mX + c)

GR&R Study

20029158 — 2023-02-09 07:39:14 UTC

Gauge: equipment or instrumentation used for measurement

R&R: (Repeatability and Reproducibility) - combination effect on operators

GR&R: determine variation amount in data linked to measurement system and comparing with variation total.

Good Linearity and Biasness

Pengiunieee — 2023-02-09 07:59:23 UTC

20029158 — 2023-02-09 08:14:03 UTC

MSA due to operator(reproducibility) & gauge(repeatability)
- MSA due to reproducibility -> involved in operator to operator variation & operator part variation

Gauge Repeatability

Pengiunieee — 2023-02-09 08:15:01 UTC

Identical characteristic and environment in a short time span

Evaluation on new digital mesuring system that measures fuel injector nozzle diameter

20029158 — 2023-02-09 08:20:43 UTC

-2 operators measure 9 nozzles collected from production floor
- each operator measured the pipes in different order twice
- 36 measurements were obtained
- specification of nozzle diameter -> 9012+/- 7
tolerance = 14 microns

MINITAB FOR GR&R STUDY (CROSSED)

20029158 — 2023-02-09 13:41:16 UTC

CONTRIBUTION

20029158 — 2023-02-09 13:44:09 UTC

Fundamental of MSA: Katelyn
Key Aspects of a MSA: Katelyn
GR&R Study: Katelyn & Qiu Ling
Root Causes for poor GR&R Studies: Qiu Ling
GR&R using Minitab(TEAM ASSIGNMENT): Qiu Ling

Identifying Linear and Bias

Pengiunieee — 2023-02-10 01:27:31 UTC

In linearity, measurement data have bigger difference compared to the reference values while bias has relatively constant and not dependent on reference value and the data measurement is higher.

Difference

Pengiunieee — 2023-02-10 01:31:08 UTC